Koen Delvaux's blog

The brilliant box camera: Zeiss Ikon Box Tengor

2015-01-24T10:01:00.003-08:00

The last thing missing in my collection of film camera's was a box camera. I was lucky enough to get this beauty for free. It is a Box Tengor made by Zeiss Ikon. According to the camerapedia, my camera is the 55/2 version, one of the latest.

It's actually pretty advanced, featuring a shutter interlock and a lock button to avoid accidental exposure when the shutter is armed. Actually, I lost 1 frame on my testfilm because I was pushing in the shutter release very slowly. Halfway through, I changed my mind and did let go of the shutter release to reframe the image, but the interlock had already engaged. The consequence was that I could't release the shutter anymore without winding on the film. My mistake.

The lens only has 1 element. I managed to clean it by keeping the shutter open with the bulb switch. When you want to focus closer, there are 2 auxiliary lenses which can be pulled into place: one should focus between 2 and 8 meters, the other one between 1 and 2 meters.

In the picture below the focus setting was 2-8 meters and the front wheel of the motorcycle was at a distance of 3 meters. The f11 diaphragm didn't give enough DOF. Only the front of the front tyre is in focus.

The picture below was also taken at the 2-8m setting, with the window at approximately 4 meters away. The f16 diaphragm gave more DOF this time, but the focal distance seems to be somewhere around 2.5 to 3 meters instead of 4 meters (=hyperfocal distance if DOF from 2 to 8 meters is to be achieved).

The 1-2 meter setting shows a similar result. The picture below was taken at f16 with the headlight at 1.5 meters from the camera. It is already slightly unsharp. The front shock absorber is still in focus, so I guess the real focal distance will be around 1m20.

The best image quality is clearly achieved without the auxiliary lenses, at the infinity setting. The picture below is taken at f16.

The camera is giving a pretty sharp image, but only in the center. Towards the edges there is quite some light falloff and a bit of unsharpness. In the image below I moved the diaphragm to f22. This increases the DOF (obviously) but it seems to give slightly less sharpness and contrast at infinity. Probably this is due to diffraction kicking in.

At f11 the image deterioration towards the edges is much more severe. Note that the image below was also taken with a yellow filter, causing a weird flare issue at the top.

I also tried to provoke lens flare by including the sun in the frame, but nothing happend. I guess flare is more of an issue with lenses with multiple elements.

I love the simplicity of this camera. It's a great experience to use one. The mirror-and-lens reflective viewfinders seem impossible to use in indoor lighting, but they are quite usable outside. The best results are achieved at f16 and infinity, although subjects can also be positioned at 1m20 or 3 meters. If you open up to f11, be prepared to use the edge unsharpness as a creative tool. f22 gives you more 'normal' images without too much darkening or unsharpness in the corners.

The shutter speed of the camera is fixed at 1/25th of a second. I measured my shutter at 33ms, so this is a more convenient 1/30th of a second.

High-res 40 Mb images are available in this Google Plus album.

BMW C Evolution testdrive

2014-10-01T01:28:00.001-07:00

A week ago I had the pleasure to test the BMW C Evolution electric scooter. A real pleasure to drive and maybe even a suitable replacement for my daily commute. Here's a video of our group riding the C Evolution around Brussels.

You have to try it to believe it: this is by far the most stable, agile, versatile and snappy two wheel vehicle I have ever driven. The electric propulsion adds a whole new dimension to motorcycling. No shifting, instant power. Always. You might not realise it, but this changes the way you drive. Even cornering seems more stable, probably due to the direct link between your wrist and the power at the rear wheel. Or it could be the very low center of gravity. Whatever it is: this machine is a joy to drive.

Acceleration is in superbike territory. Top speed is limited to 120 km/h, although I managed to top 124 km/h and another test rider managed 126 km/h. The reason for the limit is twofold: it keeps energy consumption within reasonable limits and it means BMW succeeded in lobbying to get the C Evolution certified as a 15 horsepower vehicle. This means you can drive it without a motorcycle license in Belgium.

There's only two points that might spoil the party: range is quite limited and the cost can be prohibitive. For most commuters the range of 100km will be sufficient, but in my case the daily commute is about 60km with 50km of highway. This means I will be riding at top speed most of the time, which will quickly drain the battery. It should still make 60km, but the question is what will happen to the range when the battery has a few years under it's belt. BMW guarantees the battery for 5 years without a distance limit. But I guess this guarantee will not be for 100% of the capacity.

At €15.000 the C Evolution seems expensive for a scooter, but if you buy it to replace a decent motorcycle the difference might not be so much. I drive 10.000km/year. For 5 years this means about €2.500 in gas. I guess the difference in maintenance cost can also quickly amount to €500/year. This means that if the purchase of this scooter replaces a €10.000 motorcycle, you will be break even.

For me personally it is still a stretch. I liked driving the C Evolution and I can see myself riding a similar vehicle to work every day instead of my motorcycle. But for almost exclusive highway use I'm afraid it is still a little bit wishful thinking.

First roll on my Zeis Ikon Ikonta 521/2 - lens ghosting problem

2014-09-22T00:40:00.002-07:00

This is the first roll on my Ikonta. There seems to be a problem with the lens which creates a double 'ghosting' image. This ruins sharpness.

check out Google+ photo album

Besides this, focussing and exposure are spot-on. I'll have to look into the ghosting problem. When I took the lens and shutter apart, I had to repaint the inside of the lens barrel with a sharpie. Maybe the black sharpie ink is more reflective than I tought?

Here are some illustrations of the problem. The images shown are details from a 2400dpi sharpened scan. The full image is always on the left, with the rectangle indicating the cutout. In the BMW logo, you can clearly see the the image is projected a second time on the film, only a few pixels lower. The two images overlayed created a blur. When you look at the 'BMW Brussels' text you see that it is sharp (altough it is slightly behind the focus plane at only f8) but that there is a sort of shadow below the letters, created from the second image.

The next picture is taken with the aperture full closed. Here I don't see the double image anymore, although sharpness is reduced a little bit due to diffraction. Could the problem be caused by internal reflections in the lens barrel, which are minimised when the diaphragm is closed?

The problem with the double image is that in areas with a lot of detail and low contrast it just blurs the entire image, as if the lens was completely unsharp. The image below is a good illustration of this.

I would like to solve this problem, since the Ikonta would be my favorite camera. It offers a combination of a super large negative and real pocketability. My Franka Solida (with viewfinder) doesn't fit in my motorcycle pants, but the Ikonta does. That makes it ideal to take with me on motorbike trips.

So if anyone has a suggestion on how to solve this problem: shoot!

Large format film developed in trays

2014-09-14T05:41:00.000-07:00

My first attempt at developing 4x5 film sheets was done using the 'taco method'. This method has 2 disadvantages:
- I damaged the negatives since my developing tank is not high enough (it only holds 2 reels)
- one can only develop 4 negatives at the time.

This weekend I tried to develop a few Ektar 100 sheets in development trays. The big challenge is to keep the chemicals at a constant temperature. I conducted a test for an 'au bain marie' setup. I filled a large 40x50 tray with 5 liters of water of 40 degrees Celsius. In that tray I placed a smalled 20x30 tray filled with 500ml of water, also at 40 degrees Celsius. My digital thermometer indicated that the water in the small tray cooled down to 39 degrees within 1 minute. 5 minutes later is was still 38 degrees. I figured this would be good enough for temperature control. I preferred to err on the 'too warm' side, since the shuffling of the negatives brings them in contact with the cold air, reducing the temperature more than in my test.

For the actual development I placed the bottles with developer, bleach, fix and stab in a bucket of water at 40 degrees. I filled 3 trays with water, added the smaller trays and poured the developer, bleach and fix in the smaller trays. I started a 4 minute 15 seconds timer on my mobile phone, started to count the seconds and switched off the lights. I took the negatives out of the box and when I had counted to 30 I dropped the negatives in the developer and started shuffling them around.

This was a big mistake: when you drop the negatives in the fluid, they will stick to each other. Unsticking them all took me more than a minute, causing uneven development on 2 negatives. I should have pre-soaked the negatives in the water in the larger tray, unstick them and shuffled them around once to make sure they can move freely.
When the timer on my mobile phone went off, I moved the negatives to the bleach, went into the corner of the room farthest from the negatives to switch on my mobile phone and reset the timer to 6 minutes. Then I went back to the trays to shuffle the negatives continuously and move them to the water after the timer went off. A last reset of the timer, negatives in the fixer for 6 minutes and then I switched the lights on. Whilst the negatives where being rinsed in the water I cleaned the small trays so I could reuse one of the trays for the stabiliser.

I am happy with this method. Temperature control was easy once tested, I can do a lot of negatives in 1 batch and the negatives are 100% clean without any damage.

Working with the Cambo Explorer is fun, but I still find it difficult to decide where to put the plane of focus. Sometimes I take the wrong decision and it doesn't look natural. Maybe it is best to learn by using fewer adjustments instead of trying all possible camera movements at once.

(pictures are taken at the c-mine archeological site in Genk, Belgium. All pictures in this Google+ album.)

First strobist experience in studio photography

2014-03-30T07:18:00.002-07:00

Last friday I was asked to take pictures for the 'Mobility Day' at Lampiris. Since I just received my studio kit with lightstands, umbrella's and backdrop from eBay, I decided to turn it into a studio shoot. It was only a partial success.

The setup

The theme for the day was "Mobility, it's child's play". Lampiris employees who came by bike were offered a healthy breakfast. Given the theme, I decided to photograph each participant with a children's bike, a stroller or a step. I got all of the props in my car and set off for the shoot. I arrived at 7h30. Between 8h and 8h30 about 30 people would be arriving for the breakfast. They did not know in advance that they were supposed to take part in a photoshoot.

The equipment

I used 2 lightstands, one with a big white shoot trough umbrella as the main light and a smaller one as the fill light. I also brought a tripod to mount a third flash with an aluminum foldable snoot to use as a rimlight. A fourth flash was positioned behind the backdrop in an attempt to make the backdrop completely white.
I used 2 stands with a crossbar to hold up the backdrop, which was made out of some kind of paper fabric. Since I was going to do full lenght shots, the backdrop continued under the feet of the models.
The flashes were triggered with cheap radio triggers. I mounted the Nikon 35mm AF DX lens on my D5100 and put it on f5.6/125s and shot in RAW.

The flashes I used was a Nikon SB-28, a SB-600, a Sunpak 4000 AF and a Metc Mecablitz 30.

Problem 1: backdrop folds

Unfortunately, the backdrop was supplied folded and the folds created really nasty shadows. I tried to iron those folds out of the fabric, but the fabric just melted. This material seems to be paper with a bit of plastic in it. It is (and looks) very cheap. I'll look for something else to use as a backdrop in the future.

Problem 2: background not pure white

Since this backdrop was quite translucent and since it had small holes in it, I had the idea to light it from the back. I had enough room behind my setup and pointed my Sunpak flash to the center of the backdrop, zoomed out all the way. There was still quite a lot of light falloff at the edges and even the center wasn't bright enough to blow out completely. I was already at f5.6 and didn't want to take the risk to go to an even larger aperture.

Problem 3: Rim light was too much sideways

I positioned the tripod with the snooted Metz flash right next to the backdrop. But since the subjects were relatively close to the background (I didn't have much height in the room to play with), the angle resulted in about 45 degrees sideways. This added an interesting effect to the light, but it wasn't the rimlight I intended it to be. Another problem is that the snoot was too concentrated when multiple people were in the frame (also see the next problem).

Problem 4: People don't listen

The biggest problem I had during the shoot was that my lighting and set were built up for 1 person at the time. I didn't anticipate that since people came together in small groups that they would want to have their picture taken together. The set wasn't really big enough for this. I ran out of backdrop, I ran out of flash power and I ran out of room to back up with my camera.

Problem 5: Too much fill light from everywhere

I lowered the power of the key light to have a high key lighting at f5.6 to make the background as white as possible. Then I lowered the fill light to half of that, but this was still too much. I should have reduced the fill light further, but I ran out of preparation time. Furthermore I have the impression that the white room in which the umbrella's were right against the white ceiling reflected a lot of light, essentially already functioning as a multi-directional fill light. I also had a lot of light spill from the back-illuminated backdrop.

Problem 6: I shoot faster than my flashes

After a few shots I started "doing the strobist dance". Your models are on the set, you take a shot and then need to wait 6 to 10 seconds for the next shot. This is a long time when you have 4 people modeling. So you need to keep their attention by pretending to be focusing, by adjusting their hair or generally keeping them busy until your flashes are ready again. The "just a moment, we're waiting" attitude takes a lot of spontaneity out of any shot. In reality this means most of the time your first shot will be the best one.

Postproduction

The pictures turned out OK in the end. On each picture, I had to make the background go white. I decided I didn't want to cutout the subjects in Photoshop (too much work and too artificial), but I used the adjustment tool in Lightroom with contrast to 100%, shadows to 100%, highlights to -100% and exposure to 0.6 to 1.7, whichever was needed to make the background (and the creases in it) go unnoticed without having to much difference in luminosity with other area's of the image where the background showed through (for example through the wheel of a bicycle. I didn't go as far as also painting this in with the adjustment brush.) Since the pictures had to be published a few hours later, I only made rough selections, relying on the 'smart mask' option of the tool.

Before:

After:

Conclusion: work hard or work smart

I managed to save the shoot with quite a lot of extra work. What I should have done was use 2 flashes to light the background from the front: one from the left and another one from the right. I could use the flash from the fill light for this, since it wasn't really necessary. The rim light should have been positioned in the center of the backdrop just on top of it. Maybe I even shouldn't have snooted it to accomodate for the larger groups.

I learned a lot in this experiment and I had a lot of fun doing so. I'm already looking forward to the next shoot!

Watch the whole set here on Facebook:
https://www.facebook.com/media/set/?set=a.733080866742891.1073741879.353634754687506&type=1

Canon Canoscan 9950f scanner sharpness versus Epson V500

2014-03-22T07:02:00.002-07:00

I just received this old Canoscan 9950f scanner. I planned on using it to scan 4 strips of 35mm negatives at once, but sadly the seller didn't include the appropriate negative holder.

Using the large format 4x5 negative holder I was able to scan a 4x5 negative, a medium format negative and a 35mm negative to test the sharpness of the scanner. I then compared the 2400dpi scans against the same negatives scanned with my Epson V500. I converted the negatives to black&white, since I couldn't get a decent color calibration out of the Canon scanning software. I'll have to look into that later.

As you can see, the Canon scanner delivers slightly sharper scans than the Epson V500. Note that this cannot be generalized, since sample variation between scanners can be quite high. The most critical aspect is the height of th negative above the scanning glass. I tried to optimize this with the Canon scanner, but the default height of the combined large/medium format holder seems correct. Scans using the slightly thicker Epson holder on the Canon were less sharp, as were scans made with the negative directly on the glass plate.

If I manage to get a hold of the 35mm 4-strip negative holder and I can achieve a scan without any auto-exposure using the supplied canon software I will certainly keep this scanner.

The scans below show a 100% detail section at 2400 dpi of a large format, medium format and 35mm negative. Each time the Canon scan is on top and the Epson scan is at the bottom.

First try at 4x5 large format with Ektar 100 and a Cambo with 150mm lens

2014-02-22T14:30:00.003-08:00

My first try at large format. I processed the negatives using the "taco" method, but my tank wasn't high enough, so I squeezed the negatives resulting in some visible damage. I scanned the negatives with my Epson V500, which doesn't cover the full width of the negative. This means I had to combine 2 scans in Photoshop using photomerge, which didn't went perfect. I have tone and perspective differences.

The first negative is my best. I didn't play with too many camera movements. I only used swing on the front panel to align the plane of focus with the crate besides the railway. I shot it at 1/60 and f11 on 150mm in heavy wind.

In the second image of the chapel in the field I used rise to keep the verticals. Or at least: that was the idea. I used the 150mm lens wide open at f5.6 and 1/125 since the wind was blowing head on onto the camera.

The third image is the one I like the most. I used crazy movements on the front and the back to try to make the left side of the building as big as the right side. In reality I was shooting the building sideways. On top of that, I added a front swing to shift the plane of focus completely. Only the central part of the building and the railing in the foreground in front of it are in focus. I took the image on 1/4 of a second at f32 in medium wind. It's a technical experiment with camera movements, but I like how the perspective and plane of focus turned out.

The fourth image of a toilet building in a deserted military airfield is the most disappointing. It is taken at 5 seconds with no wind (inside a building) at f22. I only used front rise to get the lines parallel. I also could have put the camera up higher, but I didn't to see how the image would turn out. There is a strange deformation in the image which I can't really point out. Maybe I forgot to put some movements on zero before framing the shot.

Denon Globe Cruiser AH-W200: great bluetooth in-ear headphones

2014-01-14T00:16:00.001-08:00

Since a couple of weeks I am the proud owner of a pair of Denon Bluetooth earbuds. They have become my daily companion when I go out running.

I am very pleased with the transparent sound. They are a little bit heavy on the bass, but for pop/rock music this is perfect. For classical music I like to tone them down a bit with the equaliser of my smartphone. They are more than loud enough. I never use them over half of the maximum volume.

The fit is very pleasant. It is a bit of a hassle to put them on, but once they are fitted over your ears they won't move. The actual earbud is attached to an extensible and moveable rod, which makes it easy to adapt them to your ears. The in-ears are delivered with 4 sizes of replacement tips. As usual, I needed a bigger one on the left than on the right. Denon also supplies a pair of Comfyfoam tips. I tried these in the past with my normal in-ears and got a very good seal, but for the Denon Globe Cruiser I don't need them. The seal I get with the normal tips is more than good enough, so I don't need to go through the trouble of rolling the Comfyfoam tips between my fingers before putting them in.

The Bluetooth connection is very stable. I read reports of people who experienced interruptions and disconnects when their smartphone was in the wrong pocket or below their waist. I have never experienced this. I even tested the range in open air. With the phone laying flat, I can walk away up to 22 meters before the connection in interrupted. With the phone in a vertical position, I even reached 34 meters.

Control of your music is through 1 button on your right ear: push one time to pause, push again to resume play, push twice to skip to the next song and push three times to skip to the beginning of the song (or to the previous song, if you're already at the start of the song).

Before purchasing the Globe Cruiser AH-W200 I also considered the Denon Exercise Freak AH-W150BK. This is the sports version of these earbuds. The sports version is supposed to be better protected against sweat and rain, but sound a bit less good. The Globe Cruiser version, on the other hand, is delivered with a carrying pouch, airplane adapters and a 3.5mm cable to connect the earbuds to a non-bluetooth device. Normally the Globe Cruiser is a bit more expensive than the Exercise freak, but since I got a good deal on the Globe Cruiser for €70 I chose this one. I think the two products are the same, but with a different styling.

Automated testing of focal plane shutters

2014-01-05T03:06:00.001-08:00

In this video I show you how I test the shutter speeds of my old analog cameras for accuracy with a DSLR. This method also allows you to see shutter bounce, shutter capping and uneven shutter speed.

After I do a series of exposures on all shutter speeds, the resulting images look something like this:

I then process these files with a Photoshop action which does the following:

cut out a part of the center of the image
apply a gaussian blur with a high radius to the cut out image
reduce the resolution and save the image

After these steps, I use the eyedropper tool in Photoshop on the thumbnails of the resulting images to quickly compare the luminosity between the reference images and the test images.

I believe this is much more accurate than using photodiodes or photodetectors. When you use a photodiode on a shutter slit width of 1mm and the diode itself has a width of 4mm, you are measuring with an error of 400%: you will think your 1/1000s speed is only 1/250.

Heartrate monitors and GPS for running: the Runkeeper-Magellan Echo dream combination that would beat Polar, Garmin, Tomtom and Suunto

2013-12-19T08:43:00.000-08:00

I did a bit of research to what option are available to track your run. Let me start by saying that if you're into equipment review there is no substitute for http://www.dcrainmaker.com/. This man has got it all covered.

1. How I started: an old Polar S210

That was 10 years ago. I bought a Polar S210 that has served me well until the battery died. I never got around replacing it, since the old T-61 coded strap is essentially a throw-away item. When I was cleaning out my closet last weak, I stumbled upon my old trusted S210 and wondered if I could fix it. I succesfully did so by cutting open the strap and replacing the battery. The watch itself was a little bit easier to fix. It just suffered from oxidised contact strips between the print and the lcd. Nothing a bit of acetone can't fix. Unfortunately I mistakenly 'cleaned' the inside of the transparant housing of the Polar, resulting in a non-transparent lens in front of the LCD. Luckily I also managed to fix that problem with carefull sanding and polishing of the plastic lens. I'm proud to have my vintage heart rate monitor back in pristine condition!

2. My Runkeeper Android + Polar Bluetooth wearlink strap

I'm a Runkeeper user since the very beginning in 2010. I like to share my runs. It keeps me motivated. But I was still missing the heart rate monitoring feature. Hence, I bought a Polar Bluetooth strap which paired with my Android phone so I could log my hear rate in Runkeeper. A disadvantage of this solution is that the Bluetooth straps eats about 1 battery very 10-15 hours. But the real dealbreaker is that I can't see my heartbeat in realtime, since my phone is attached to my upper arm.
Mind you that this heart rate strap is not the same as the Polar Bluetooth SMART strap. The latter uses a newer low energy Bluetooth technology which is only available in Bluetooth 4.0 compatible devices under Android 4.3 and on Iphone 4s and 5.

3. Let's look around for a heart rate monitor with built in GPS

My refound old Polar S210 sparkled my interest again in a heart rate monitor. I looked around what is available on the market of combined GPS + heart rate monitor watches and came up with the following list:

€169 Garmin Forerunner 210: allows for programmable interval training + shows average pace/heartbeat per interval. Does not have a 'virtual training buddy'. Easy sync to Runkeeper via Garmin Connect + http://www.copymysports.com/
€219 Polar RC3: is fully programmable, has a virtual training buddy and has a neat arrow pointing into the direction 'home'. Sadly, it doesn't show average lap/interval pace/heartbeat. It should be possible to import the Polar files into Runkeeper via a computer, but it will not be plug&play.
€219 TomTom Runner: fully programmable via the TomTom website (great!), many customisable data to display (also for laps and intervals), has a virtual buddy feature but no possibility to export or sync data to Runkeeper. Extra feature coming up in the Tomtom software: cadence information without a footpod, using the built-in accelerometer. Files can be exported from the site to be importer in Runkeeper. It remains to be seen if an automatic synchronisation service will emerge.
€299 Garmin Forerunner 220: This successor to the FR 210 offers many advantages: a color screen, fully programmable via the Garmin website, programmable data screens including lap/interval information. No virtual training buddy, but at least pace alerts are offered. Like the TomTom it also has an internal accelerometer for cadence information. And the biggest advantage: you don't need a computer to upload your runs, since it can sync automatically over Bluetooth 4.0 with the Garmin app on your smartphone (which then can automatically sync it to Runkeeper with http://www.copymysports.com/)

When reading about the Forerunner 220, I realised that all of the other watches require you to find a data cable, find a computer, hook them up and press several buttons to upload, copy and sync. Not something I am looking forward to do after each run. So for me, the only remaining option seems to be the expensive Garmin Forerunner 220.

4. The ideal (fictitious) solution: Runkeeper + Magellan Echo

When I looked a bit further, I realised the whole paradigm of stuffing intelligence in a watch is completely flawed. Our smartphone provide a much better user interface to configure things, they have better sensors and the innovation cycles happen in freely downloadable software instead of in expensive hardware that is difficult to update.

I found out about the freely programmable intervla trainings in Runkeeper. I found their training programs. I found the RunSafe app which not only records your cadence using the accelerometer in your smartphone, but also advises you in realtime how to adjust your stride. And the appmaker already made sure it can coexist with Runkeeper. I also realised that I like to listen to my music when I go running and there is not a single GPS watch that can play MP3 files.

The solution is simple: applications and functionality extensions belong on your smartphone. The device on your wrist just needs to be a dumb screen with buttons on the side. Exactly that is what Magellan created with the Magellan Echo.

The device only came on the market a month ago and there are already 3 apps that are compatible with it: the Wahoo fitness app, Strava and iSmoothRun. The last one is really complete and allows you to create custom interval trainings. It can also upload your runs to Runkeeper, but sadly it is only available for iPhone and not for Android.

Conclusion

Smartphone based solutions will eat the running watch market and Runkeeper is very well placed to grab a big share of it. If they only would want to create an integration with the Magellan Echo, I would be one hell of a happy Runkeeper Pro subscriber!

UPDATE 23/12/2013: I decided on a Garmin Forerunner 610

I went running on an athletics track today and the disadvantage op a smartphone based solution became immediately evident: the accuracy sucks. Since there is no Android support (from Runkeeper or others) for the Magellan Echo anyway, I will buy myself a Garmin Forerunner 610. This is a top of the line sportswatch which is heavily discounted now that the 620 has been released. I will sync it wirelessly towards a Netbook that is serving as a server in my home and hence is always on. So when I return from a run I just have to click 1 button on the Netbook and the run will sync automatically.

Here is a screenshot from what Runkeeper captured on the athletics track:

Black & White battle: Samsung Galaxy 4 Zoom versus Leica CL with Kodak T-max film

2013-10-27T01:02:00.001-07:00

I like to shoot film. There's something about the process. The cameras. The anticipation. The wait. Shooting film makes photography fun for me.

Even in 2014, film photographers still believe they can get a look with film that is not possible with digital. I can guarantee you that the images from my Kodak Porta and Ektar colour negatives have more intrinsic qualities to them then my Nikon DSLR files. Not because I say so, but because people systematically choose my analog pictures over my digital ones.

The test

But is this also the case for black & white? I decided to take the test with a roll of Kodak T-max 100 (TMX). This is the finest-grained black&white film and should therefore compare well with digital. I also could have chosen a film with more 'character' like TriX, but that would make any comparison with digital pointless, since it renders the images in a vastly different way.

The digital camera I will be using for the comparison is not my Nikon DSLR but my newest camera: a Samsung Galaxy S4 Zoom smartphone, which was given to me with the request to put it through it's paces. The S4 Zoom has a large sensor (for a smartphone: 1/2.33'') and a 10x Zoom. Read my review for more information.
For the test I shot the T-max with my Leica CL through an Industar 55 lens. The 24 zoom lens of the Galaxy S4 Zoom was used at 2.3x magnification, which corresponds more or less with the same angle of view as the Industar. The Lecia was metered with the built in meter and the S4 Zoom was put on auto.

The images

The results blew me away: the Galaxy S4 zoom could render the same black & white tones as the T-max film. I processed both the scans of the film and the JPGs of the smartphone in Lightroom, but I didn't touch the colour-to-B&W mixing for the smartphone pictures.

There are differences in rendering and exposure, but not as much as I had thought. The Galaxy S4 manages to save both highlights and shadows in most images. Even the clouds where salvageable in Lightroom.

One problem area is DOF. The small sensor of the camera (compared to the fullframe, well: film) does not allow to render objects unsharp. The difference can be seen below.

As you can see in the 100% crop below, in general the smartphone delivers cleaner images with more detail. Only on the shadow side (to the right of the vertical strut) we see that the smartphone's noise reduction has erased all textures. We also see some noise in the smartphone image, compared to a little bit of film grain in the T-max image. The general impression of the image is still more pleasing with the T-max than with the sensor.

The most interesting difference is in the last picture, where there is less detail in the sky in the S4 Zoom image as in the T-max image. Maybe this has to do with the colour mixing, since this image was taken at 1/8s after sunset. The sky was coloured blue and orange.

The full resolution images of the whole film are available in this Google+ album. The images from the Galaxy S4 zoom are here on Flickr.

Personally I still prefer film. Not because it gives me a look which I can't recreate with digital, but because of the pleasure it gives me when shooting. The limitation of only having 36 frames available makes you think about what you are shooting. During this test, I was shooting as I would with film and then took a second image with the digital camera. I don't think that without the film camera and the limitations it offers me I would have come home with the same images.

So will I keep shooting black & white film? Hell, yeah! There are many ways to an end and my way to create images is with a classic film camera. It's as simple as that. Up to you to see what floats your boat.

Samsung Galaxy S4 Zoom in-depth camera review

2013-10-18T03:53:00.000-07:00

This is a follow-up review of the Samsung Galaxy S4 Zoom, which I already compared with 2 other cameras. In this review, I'll look at the hardware, software and (most importantly) what the camera delivers in real life daily usage.

All of the images in this review are available in their full resolution in this Flickr album.
http://www.flickr.com/photos/kodel/sets/72157636663554624/

Hardware

Most reviews split this out over chapters, but I'll handle everything at once. Besides the camera, this phone is a mid-market device with a lot of technology inside. It has the usual compass, acceleration and position sensors, GPS, Bluetooth and WiFi, but it also features extra's like a light sensor, infrared blaster and NFC. It's also nice to have an SD card slot in case the 8GB of built in memory runs out.
It is, however, slower than higher end phones. This doesn't bother me so much, but I would happily pay €100-200 more if the phone was as fast as my Nexus 4. For everyday usage it is certainly fast enough.
The screen might have a slightly lower resolution than bigger phones, but I really like how vivid colours look on it. Samsung even provided an application that automatically lowers the saturation of your screen when you start the Gallery app, so you will not think your pictures are over-saturated. Nice.
A very positive aspect is the sound quality. I have always felt that very thin phones tend to sound kind of, you know: thin. This telephone gives a rich and powerful sound for which you don't have to position your ear exactly over the speaker.
The battery could be better. I can't make it through a day of heavy smartphone + camera usage on a charge. Given the bullk of the phone, one would expect a larger battery.

Stock Android versus Samsung applications

Like most smartphone manufacturers, Samsung delivers it's phones with a modified version of Android. This is not because their version is better than Google's, but for two reasons. The first reason is that marketeers want something to put in their advertisements. Therefore, each phone comes with bunch of custom applications and a custom user interface shell. The second reason is that device manufacturers receive money from application builders to include applications as a default option on their phone.

There is nothing wrong with this. It also happens with laptops, where manufacturers bundle software (called 'bloatware') with Windows. The difference with smartphones is that you can't erase the extra functionality. There is no easy way to go back from the Samsung email application to the stock Android email software. I find the Samsung application to be less user friendly than the stock application. One example: when I want to view my 'unread mails' it takes me 2 clicks on the standard Android email application and 3 clicks + 1 scroll on the Samsung one. It also doesn't have a spellchecker when the predictive text feature of the Samsung keyboard is deactivated.

There is a solution, however: a young company called Cyanogenmod, originating from the hacker community, provides stock android versions and installation procedures for most phones. My decision to keep using the Galaxy S4 Zoom (or pass it on to my wife) will be based on the availability of Cyanogenmod for this device.

Camera: quality

I can't stress this enough: this camera is better than anything you've seen from a smartphone. There is a real quality difference between camera phones and compact camera's because of the larger sensor and the better lens. This device has both. Joy! Really. Even for casual family snapshots you'll notice the difference.

Camera: zoom

The 10x zoom is the primary sales argument of this camera. Personally, I don't use a lot of zoom in picture. (I still use ancient analog camera's with prime lenses). But most people will have a blast zooming in. Below is an example of just how much closer you can get without using your feet (24mm-240mm)

Camera: flash

Aha, a real flash! It makes a vast difference for dark indoor shots. The camera makes a good effort to balance the flash light with the environment. Even at the 24mm wide angle the image is quite evenly lit. Pay attention to large bright surfaces close to the lens. In the image below I had to add +1 stop of exposure compensation. Otherwise the image was too dark.

Camera: tripod

Not enough light? With a smartphone this means you're stuck with shaky and noisy high iso images. Not so with the S4 Zoom. You actually have two options: just take the picture and relay on the excellent image stabilisation coupled with the low noise high iso performance thanks to the bigger sensor. Or you can mount the S4 Zoom on a small pocket tripod. The first picture is in auto mode, where the camera chose to keep the diaphragm open (obviously) at f3.1 and applied an ISO setting of 1250 and a shutter speed of 1/6 second. The second picture is with the camera mounted on a tripod combined with a manual exposure of 10 seconds, a closed diaphragm of 8,8 and a noiseless 100 ISO. I like to have options and this camera certainly offer them.

Camera: portraits

So I told you I wouldn't use the zoom. I was wrong. The one thing that smartphones are bad at is portraits. The reason is simple: they all feature a wideangle lens which is good for general snapshots. But for portraits you need a longer focal lens (read: zoom in) to render the proportions of the subject's facial features correctly. The S4 Zoom offers you just that. In the picture below I zoomed in to about 77mm, about the ideal focal length for portraits. Skin tones are rendered naturally, even under halogen or fluorescent light (as far as possible). The white balance is also spot on, even in difficult circumstances like the yellow sweater below.

Camera: backlighting

Cameras are easily confused when the scene is does not correspond to your typical average 18% gray reflectiveness. This is best seen in images taken where the camera is held in the direction of the light. I was pleased to see that the auto setting dealt with this backlight without a problem. The exposure is just like I would do it manually.

Camera: built in editor

Samsung did provide a fancy built in Photoshop-like piece of software. You can draw over parts of the image to select them and then locally adjust contrast, brightness, etc. I tried it on the image below. The face close to the camera is nicely lit, but the group of photographers on the left is too dark. I tried to lighten them using the built in editor, but it did get a very blocky result. There is no feathering at all applied to the selection. Not useable.

Camera: malleability of images with Snapseed

After the fiasco with the Samsung editor I tried another editor. In the spirit of this all-in-one device, I didn't switch to Photoshop but I used Snapseed. This Google-owned editor is simply amazing! It has it's own particular user interface and workflow, but try it and you'll never go back. I can see myself skipping photoshop for many images using this editor.

Anyway, back to the S4 Zoom: what I want to show you is that the JPG images of this camera have a lot of room to change the appearance of the image in postprocessing. With other cameras you sometimes see JPG artifacts or lack of a continuous tone, especially in the highlights. It would have been better if the S4 Zoom would output RAW images for later postprocessing (like a professional camera), but the JPGs are certainly very usable in Photoshop.

Below is an example of a fragile image that has been postprocessed (to black and white) with Snapseed.

Camera: image format

As you can see in the previous image, I like to crop my images to the 3:2 format. It took me a while before I realised this, but the Galaxy S4 Zoom does have a native sensor of 4:3 at a resolution of 4608x3456. The standard setting however is 16:9, giving you only 12Mpx of the 16Mpx delivered by the sensor. Samsung probably does this to match the image aspect ratio to the screen of the S4 Zoom, but once I found this out I always kept the camera in 4:3 format. It's better to keep your options open and crop afterwards if you want to.

Camera: macro mode

Close-up pictures with this camera are amazing. You don't even need to switch to the macro program. Just make sure you're zoomed out all the way. You can get to within centimeters of your subject. Below is an example of keyboard of my Macbook. I focussed on the letter "G". If you download the full resolution image from Flickr (see above) you will see that every pixel is tack sharp. A great performance.

Camera: Smart modes

I'm not good at Smart modes. I never seem to understand the decisions the camera is trying to make for me. Food mode gives saturated colours, landscape mode pops the blue, etc. I prefer to do these things afterwards in Photoshop. There are however a few modes that are fun. "Best Photo" automatically picks the sharpest image out of a series. "Best Face" allows you to take several pictures of a group of people and to pick for each person on which picture his or her face is the best. The camera will then combine the different faces in 1 picture. "Animated Photo" makes a short videoclip but lets you paint an area that you want to "freeze", so you get a picture with only 1 area in motion. "Eraser" takes multiple images of the same scene and lets you erase subjects that move through the image. Good when you want to photograph the San Marco square seemingly without anyone in the image (but again I would advise Photoshop for this). The results of these intrusive Smart Modes is quite good, but not perfect. If you want to have fun and share with friends, they will be perfect.

Below is an example of a scene that was lit wrong by auto mode. The image was much too dark. It is a typical example where you want to apply exposure compensation to keep the detail in both the sky and the tree. To see how an amateur would cope with the Smart Modes I did what an amateur would do: I switched to landscape mode. The blues and greens are a lot more saturated by this mode, but the sky is not recognised as such and the camera still underexposes. In the second image I exposed manually the way I feel the scene should be metered. I didn't increase the saturation in blue and green yet. I think I will apply a bit of dodging and burning in photoshop first. It is clear to me that the manual mode is still a must. The Smart modes are fun, but not useable for more serious photography.

In comparing the images, I also noticed that the S4 Zoom never closes it's diaphragm in Auto mode. It just stays open and increases the shutter speed, even to unnecessary high values likes 1/800. I don't know why Samsung programmed the camera like this. Closing the aperture above a certain shutter speed (e.g. 1/60) would give much better images.

Camera: video

I don't make a lot of videos, but the one I tried (in low light) turned out good. The sound quality is also good, despite there being a lot of background noise in this performance of "The Cup Song" by my daughter.

Conclusion

This smartphone is a wonderful device and a bit of a missed opportunity at the same time. Wonderful, because it is the only device in the world to give you compact camera photo quality together with a very usable mobile phone. A missed opportunity since this nice product could be so much more with a plain vanilla Android, non-zoom lens but RAW output in return and a decent UI for the manual photography. I imagine it would appeal to Fuji X100s shooters who would happily fork out 1.000$ for such a device.

Now it is still a nice product, since most people are happy with the existing camera in their phone, but it aspires to attract a wide audience. It might be exactly this "happy snapper" implementation that might make it miss it's nice.

I think I will keep the phone, unless we see something like a Nokia 1020 appear in the Android market. If you love photography, the Samsung Galaxy S4 Zoom is made for you!

PS: After writing this review I discovered a dust particle inside the lens. This occurred after carrying the camera in my pocket for only 1 week. It's is very much diffused into unsharpness at the wide setting, but when you zoom in renders images unusable. Check out the effect here: http://www.flickr.com/photos/kodel/10333347925/in/set-72157636663554624 If this issue doesn't cure itself or it reappears it will be a definite dealbreaker to keep the camera after my 4-week trial period.

Pixelpeeping Test: Samsung Galaxy S4 Zoom, Google LG Nexus 4 and Panasonic FS-40 compared

2013-10-12T13:09:00.001-07:00

Since two days I'm the proud owner of a Samung Galaxy S4 zoom. The device is a marriage between a smartphone (the S4 mini) and a compact camera. As such, you can finally use a decent lens, a larger sensor and an optical zoom on a smartphone. I'll review the device as a smartphone in a later blogpost, but this is already my evaluation of the image quality of the camera.
Samsung seems to use the 10x zoom as the primary sales argument for the S4 zoom. I'm not too much interested in zooming myself. I even sold the 70-300 zoom lens for my DSLR since I rarely use those focal lenghts. Most picture are wide to normal, in a range of 24-50mm. This is why I'm very excited to see a wide 24mm lens on the S4 zoom. It's not really the same diagonal field of view as a 24mm full frame camera, since the S4 zoom delivers a 16:9 image instead of the classical 3:2 format.
For this test I did not try the zoom and I focussed on the following aspects:

pure image quality: sharpness, distortion, color rendition
exposure, contrast and color balance. You can adjust these things in postprocessing, but that makes it tedious.
focussing: does the camera focus correctly after indicating the correct point of focus.
jpg quality: since all 3 camera's lack a RAW mode, the built in software applies a tone curve, sharpening, noise reduction and (for the Panasonic) even a stretch function to minimise lens distortions. I'll evaluate if the files are not over processed.

I decided to test the smartphone/compact against a smartphone (my current Google LG Nexus 4) and a low end compact camera (my Panasonic DMC-FS40). This test is without any image post processing. I already did a test to compare the Panasonic against the Nexus 4, which gives different results.

Below you will find 6 test images. Each image represents a specific challenge. If you want to download all full resolution images, download them from Flickr. Note that during this review I am comparing on 100% size, which is showing the individual pixels captured by the camera's. This is not really fair, since the camera with the least pixels (8Mpx Nexus 4) will appear to be on equal foot with a camera that has more pixels (16Mpx). In reality, if the images at 100% look equal, the 16Mpx will show twice as much detail as the 8Mpx camera. To further add to the confusion the 24mm wide lens of the S4 zoom captures a wider image. The result of this is that the higher pixel count of this camera is spread out over a broader scene. The consequence is that the 100% cutouts will show about the same image magnification, despite the higher pixelcount of the S4 zoom.

You will each time see an overview comparing the images from the 3 camera's, followed by one or two detail images. The Nexus 4 is always to the left, the S4 zoom is in the middle and the Panasonic is on the right.

1) outdoors, good light (warm morning sunlight)

This is the easiest shooting situation. Almost every camera can produce a good image in these circumstances: iso can be kept low, shutter speed can be high so there is no camera shake and the diaphragm can be closed. I adjusted these settings manually to get the best possible image quality out of the cameras.

We see that the Nexus 4 looks the sharpest. This is because it applies quite heavy sharpening. The Panasonic looks unsharp. When you look at the bricks, you see that too much noise reduction is applied, even at it's lowest iso setting. The S4 zoom shows the most detail on 100% enlargement. It looks a bit softer than the Nexus 4, but this is because the sharpening has been kept within more reasonable limits. It does render the scene too yellow, however (note that this was in warm morning light).

The corner detail image is interesting: it appears the S4 zoom is very unsharp here, but this is due to a lower depth of field thanks to the larger image sensor. That is a bit surprising, since I used the smallest diaphragm (the S4 zoom only has 2) and a broader field of view. But apparently the sensor is that much larger than the one in the Nexus 4 that it is still having a lower depth of field. This is actually a good thing when you want to zoom in (that's what the camera is intended for, after all) and isolate a subject or take a portrait.

2) outdoors, medium light (rainy afternoon)

In this image, we already see some more differences appear. It becomes even more apparent that the Panasonic is soft. Even more noise reduction has been applied by the camera to battle the high sensor noise at this higher iso. It is clear now (look a the pavement) that the S4 zoom has more sharpness than the Nexus 4. There is also more detail, as you can see in the streaks the rain left on the hood of the car.

I added an image of the corner again. The lamppost still falls within the depth of field. You can see that this time the corner sharpness of the S4 zoom is higher than the Nexus 4. This proves that the unsharpness in the first image was due to DOF. The Panasonic shows some chromatic aberration on the lamppost, even tough there was no sun to be seen (conditions were very overcast).

3) indoors, very low light

Ouch. This is where all 3 camera's show their limits. The light was very dim, pushing the iso settings all the way to the top. The Nexus 4 shows more image noise than signal and definately lost here. The Panasonic holds up well this time, masking it's unsharpness behind the noise. The S4 zoom wins out, showing the best colour saturation and contrast with the least amount of noise. Due to more built in softening (to mask the noise) the sharpness is no longer optimal, which is to be expected in these conditions.

4) indoors, very low light with flash

This image is taken in the same bad conditions, but with flash. Although the Panasonic has a decent flash it still decided to use a high iso setting. The result is a well-lit, but grainy image. The LED flash of the Nexus 4 doesn't cut it at this distance of 3 meters. The image is a lot better than the one without flash, but sharpenss is gone and noise is all over the place. The Galaxy S4 zoom gives a very good image. No colour casts, no washed out highlights and still a nice balance with the background. It remains of course a flashed picture, which always kills the mood a bit and can give red eyes. But remember you can always take the same picture without flash with the S4 zoom, even in the worst indoor lighting conditions.

5) outdoors, bad light (just before sunrise)

The 3 camera's show similar exposure, contrast and colour rendering in this image. All 3 manage to save the mood of the scene. There is not much light, but even the low amount of light just before sunrise is enough to keep all of the 3 contestants out of the danger zone. Again, the Panasonic is softer, the Nexus 4 applied more sharpening. For the S4 zoom I zoomed in a bit and it seems like this was enough to make the S4 zoom lose it's slight lead in pixel-per-pixel detail. I'll evaluate the impact of the zoom further when I write the full review of the Galaxy S4 zoom.

In the detail of the roof and the border of the chimney with the sky we see that the Nexus 4 exhibits some chromatic aberration. We also see how much unsharpness the Panasonic shows outside of the center. Not only does it have a bad sensor, but it also seems to have a bad lens. The S4 zoom shows almost no chromatic aberration.

6) macro, good light

The S4 zoom is clearly the best at macro distances. The other two camera's were taken as close as they could focus, showing their highest image magnification. The S4 zoom could go so close that the lens would shade the whole image, but I didn't do this in order to have an image that we can compare to the other camera's. It can actually focus to about 2 cm from the lens. Strangely enough, it can only do this at the widest zoom setting. If you zoom in, the image magnification that is possible soom becomes much less. Nevertheless, this is an impressive macro performance of the S4 zoom. The diameter of the puppet is about 4cm.

Conclusion

The Galaxy S4 zoom is a clear winner. Low noise, good sharpness and detail. Exposure, contrast and saturation are about where I like them (not over the top). Image sharpening and noise softening are also within reasonable limits to still allow for postpocessing. High iso's perform better than on smaller sensors.

It's doesn't produce DSLR quality and an advanced compact will beat it's image quality. It is about on par with a medium level compact camera. The Panasonic DMC-FS40 performs bad. It's images are noticeably unsharp and image noise pops up quickly when the iso is raised. The Nexus 4 shows it has a decent smartphone camera, provided the lighting is right and you might want to bump saturation and contrast in postprocessing.

I remind you again that this review is comparing camera's on a pixel-by-pixel basis. The Galaxy S4 zoom has more pixels than the other camera's, so it will give you much more detail when you look at the whole image file.

Learning how to do dodging & burning in Lightroom

2013-10-08T12:14:00.000-07:00

I normally don't do doding and burning. But if you want to get good results from black&white, you have to learn how to dodge & burn. Instead of learning the skill in the analog darkroom, I decided to start in the digital darkroom with Lightroom.

This is my first attempt at the skill in Lightroom. I probably went overboard on some pictures. I also used localised clarity, contrast and sharpening. The extensive controls in Lightroom make it easy to manipulate parts of the image.

Below are the 12 pictures of my roll of TMX that came out of my Rolleicord. On the left you see the picture with only a curves adjustment, on the right you have a fully processed image.
The full resolution 25 Mpx images are available on Flickr and on Google+

Feedback is welcome!

Calculate exposure time in the darkroom for different enlarger heights

2013-10-06T09:48:00.000-07:00

When you do your own printing as a photographer in the darkroom, you have to make a lot of testprints to determine the correct exposure. The problem is that wasting paper on teststrips is expensive.

One way to save on paper costs is to make your teststrips and test exposure on a small image format and then scale up the image once you are happy with how your print looks.

There are 3 ways to scale up your image:

make new teststrips to match the previously found exposure (which is wasting paper again)
use a lightmeter to find the difference in light output between the new and the old height of your enlarger
calculate the new exposure times

I opted to make the calculation by using some simple trigonometry. Maybe others can benefit from this as well. If you don't want to mess with the formula, you can use this online spreadsheet to mark your own exposure compensation f-stop setting on your enlarger column.

Here are the instructions:

Turn the enlarger all the way up and insert the negative carrier without a negative. Focus on a piece of paper under your enlarger. Now measure the longest side of the projected image on the paper and note this value as "y".

Then turn the enlarger all the way down (or: the lowest level you might ever want to use), focus and mark this point (or write it down if your enlarger has a measurement scale on the column). Measure the the longest side of the negative carrier mask on the paper again (ie. the width of the projected image) and note this value as "x"

You now have two marks on your enlarger column. Measure the difference in height between the lowest and highest mark and note this value as "c".

Now we are going to calculate the light output for the differences in enlarger height in f-stops, starting from the highest position (=lowest light output) as reference value "1".

To find the F-stop value for a given height "h" (in centimeters above the minimum height) use the following formula:

f-stop= (y / (x+ h/c * (y-x)))ˆ2

(or, use the online spreadsheet)

Here's an example of how to use this:

My values measured c=100cm, y=50cm and x=5cm.

When I make my teststrips at a column height of 25cm above the lower mark and I want to make my final print at the maximum hight (to make a 50cm wide print), I need to multiply my found exposure time with 9.47.

Smartphone camera and zoom compact compared: Nexus 4 vs. Panasonic DMC-FS40

2013-10-01T09:35:00.003-07:00

My daughter uses a cheap Panasonic compact camera. It is a 14Mpx DMC-FS40. Just for fun I decided to compare it against my LG/Google Nexus 4. I was surprised to see how good the phone held up against the compact camera. To see the rest of the images I took with the Panasonic, check out the Google+ album.

The Nexus 4 is on the left, the Panasonic on the right. Without processing in Lightroom, the color balance is a bit too cold.

Since the picture is taken in warm evening light, I adjusted the color balance on both images to reflect this and I set the blackpoint and whitepoint, which weren't too far off on both originals.

Below are 100% crops of both images compared. You can see that the straw is a bit sharper on the Panasonic. The hair in the Nexus 4 image also suffers quite a bit from noise reduction.

In the detail of the chair at the table behind us it appears the Panasonic has more depth of field. The Nexus 4 took the shot at 4,6mm at f2,65 and 1/40s and the Panasonic used 5,7mm at f2,9 and 1/60 sec. Since the angle of view is about the same, we can conclude that the Panasonic has a bigger sensor and the depth of field should be smaller at the same angle of view. Interesting. I assume the difference in the sharpness of the chair is a combination of less noise reduction and the sharper lens of the Panasonic.

The glass clearly shows that the Panasonic is sharper. The red around the letters is also rendered much better.

The biggest difference is noticeable in the green trees in the background. The sharpening + noise reduction of the Nexus 4 ruined this area of the image, whilst it holds up fairly well on the Panasonic. Also notice the difference in sharpness of the hair.

The jacket does not show that much difference (sorry for left/right reversing the images in this comparison: the Nexus 4 is on the right this time). There could have been subject motion on the panasonic image or there could be a difference in focal points.

Conclusion

The compact camera is one of the cheapest you can get (at around €60) but still it offers an improvement in image quality over the Nexus 4. If you zoom out it also has a wider field of view than the Nexus 4, which is definitely a plus. So it still pays of the use a compact for day to day snapshots.

I actually got interested in how a compact camera performs since I registered for the #GalaxyS4ZoomBE test in Belgium: 500 people get to test the Samsung Galaxy S4 Zoom, which is a hybrid smartphone/compact camera. The camera part should exceed the quality of the Panasonic I used, so that will become an interesting test!

Polaroid land camera with Fuji FB-3000B

2013-09-22T00:30:00.001-07:00

I tried out the Fuji FB-3000B instant peel-apart film in my Polaroid 320. Not so excited. I prefer the 100c.

I got several image artifacts (light leaks in the pack or camera, maybe more than with 100c due to the higher sensitivity?). Exposure was also way off: the first 2 pictures were underexposed by 2 stops. When I increased exposure, the next pictures were overexposed. In the end I settled on no exposure compensation for the brightness control, which was about right (but 2 stops under at the start). Strange.

The images shown here are scans of the negatives. In contrast with the 100c negatives you can't wash & bleach these, so you have to scan them as reflective material. This makes you lose a lot of detail in the shadows. The highlights are preserved better and the images are also sharper than what you would get if you just scanned the picture itself. The biggest problem is keeping the negatives from touching anything (especially dust!) whilst they are drying.

The rest of the images is available on Google+

Creating negatives from Fuji FP-100c polaroid peel apart film

2013-08-31T14:14:00.000-07:00

Yesterday I took my first Polaroid picture ever. I used my Polaroid 320 land camera and a pack of Fuji FP-100c. It was pure magic to see the image appear in front of my eyes.

What disappointed me was how soft the image was. So I decided to see if I could salvage the negative, which is rumored to be much sharper. It turned out to be very easy to do.
Step 1: tape the negative (emulsion side down) to a piece of glass

Step 2: drip some household bleach on the black backside

Step 3: wiggle the glass to distribute the bleach on the negative

Step 4: due to surface tension, the bleach will not distribute evenly. Give it a hand by moving the bleach around with the edge of a paper towel

Step 5: wait 2 minutes (I used a bit too much bleach here)

Step 6: rinse under the faucet

Step 7: dry the back and remove the tape, hang to dry

As an alternative to step 7 you can also wash the emulsion side with cold water. This will remove the developer (goop) wich is still on the negative. I learned that this step is unnecessary and dangerous. If you make sure the emulsion side doesn't get well at all, the goop layer will have no impact on sharpness and only a minor reduction of contrast, which can be remedied in Photoshop afterwards.
The danger of trying to wash the goop of is that you can damage the emulsion.In any case, I would never touch the emulsion side with anything (Remember that paper towels are very abrasive. You will cause scratches).

Below is a comparison between a scan of the negative and a scan of the positive Polaroid picture itself. You can clearly see that the negative is much sharper then the positive. In this case, a bit of bleach got under the corner of the negative. Some people do this deliberately to create an artistic effect.

(full res pictures available on Flickr)

Negative:

Positive:

Here is a detail of the negative (left) compared with the positive (right).

The negative below had the goop washed off it's emulsion side. You can see that there is no difference in sharpness with the negative above.

Negative:

Positive:

If you want to create negatives like this, it is important to shield your negative from light just after development. Whilst the negative is still wet, the developer stays active and the negative will get darker and fogged when exposed to strong light. I still need to experiment a bit to learn how big this effect is, but I already experienced that the dashboard of a car in full sun is not a good place to let the negatives dry.

Mamiya 528AL: worst camera ever made

2013-08-20T03:07:00.003-07:00

I recently picked up a Mamiya 528AL for a couple Euro's at a flea market. My curiosity was triggered by the fact that it was an SLR, but with a leaf shutter.

The camera didn't function, but I managed to repair it. It was the most complicated camera repair I ever attempted. The mechanics of the camera are very complex. To expose an image the following steps need to be executed in the correct order:
- the shutter needs to close (since it's an SLR it is open in the normal position)
- the mirror needs to flip up
- the diaphragm needs to close
- the shutter needs to open
- the shutter needs to close
- the diaphragm needs top open
- the mirror needs to flip down
- the shutter needs to open again

This whole sequence is commanded by a ring around the shutter assembly with various nooks and crannies. If anything doesn't work smooth, you miss an exposure: the mirror will flip up and down, but the shutter will not have closed and opened, resulting in a blurry 10x overexposed image.

In the end I got it working again. Even the light meter for the aperture priority mode and the flash are functional again. But when I ran it's first roll through it, I was very disappointed. The camera is rubbish. The lens is no better than what you get on a cheap plastic compact. In fact, I have taken razor sharp images with a plastic Yashica MF-3, but the Mamiya is only good for dreamy (euphemism for unsharp) portraits.

Most of the images on my roll of 24 were either badly overexposed or unsharp. I saved only 12 usable images from this roll. In fact, I didn't even bother to wait for the negatives to flatten, hence the moiré you see in some of the images (from the negative touching the scanner glass).

The rest of the images is available on Google+.

Scanning black&white negatives: sharpness gains in using individual channels

2013-08-15T05:53:00.000-07:00

Some people claim it is better to scan a black&white negative in colour mode and only use one of the 3 color channels. The reasoning behind this is that overlaying the 3 different scans of red, green and blue will be less sharp than only using the pixels of 1 channel.
Scanning in colour gives you the option to separate the 3 channels in photoshop and work with only 1 channel to create your image. I tested this theory with the image below. The image is shot on Kodak Tmax 100 and stand developed in Rodinal 1:100. Sharpness is not that fantastic due to the limited sharpness of my Epson v500. Even the very basic lens of the Franka Solida I that took this picture is outresolving the scanner.

In the detail below (100% crop of a 25Mpx scan) I compared 4 images: a black&white scan (using the 3 colour channels) and an image using only red, green and blue channels. I noticed a slight shadow clipping in the blue channel and an even smaller highlight clipping in the green channel. The red channel wasn't even filling the histogram, showing the most headroom.

Resolution-wise, I don't see a difference in the 4 images. Maybe the black&white scan (RGB image) has a slightly better tonality, but that's it.

Conclusion: it doesn't make sense to scan a black&white negative in colour mode and separate the different channels. Just scan your black&white negatives in monochrome mode.

Filling engraved letters with black paint

2013-05-23T13:37:00.002-07:00

I bought a Nikon EL2 the other day. The camera was in good condition, but the black paint in the Nikon engraved brand had chipped off partially. With a toothpick I removed the rest of the paint.

To fill the engraving with paint, I used a toothpick to apply tiny drops of ordinary black satin alkyd paint. I made sure the paint spread all over the engravings.

The difficulty is in removing the excess paint. When you try to wipe it with paper, the paper will suck the paint out of the engraving. The trick is to wrap a Kleenex tight around your vinger and then move with a very quick and light pressure swipe over the area to be cleaned. I ended up doing this several times and I stopped at the result in the middle of the picture below. If I tried to go any further, I had to reapply paint, which ran all over the faceplate again.

After letting the paint dry for an hour, I removed the traces of paint using a Kleenex lightly dampened with White Spirit. Again in quick, low pressure motions. Avoid getting any White Spirit in the embossed area.

I'm happy with how it turned out. In the right part of the capital letter 'N' there is a small blemish in the paint, caused by a dust particle that got caught in there whilst the paint was drying. I'll leave it this way to avoid removing the paint.

The EL2 is looking lovely with the logo restored to nice matte black!

Reflecta Proscan 7200 test & comparison with Epson v500

2013-05-09T01:34:00.002-07:00

I used to scan my negatives with my trusty Epson v500. Now that I am more and more getting into 35mm instead of 120 film, I am a bit disappointed by the lack of sharpness in the Epson v500. After reading many reviews, I decided to order the Reflecta Proscan 7200.

My first tests are not so positive: I can get the same resolution with my Epson v500 as with the Proscan. Also, the Epson v500 seems to have a higher Dmax than the Reflecta. It turns out my 35mm sharpness problems were caused by curved negatives in the flimsy Epson film holder and not by a sharpness problem of the scanner itself. So as a first step in improving the quality of your 35mm scans on an Epson flatbed, I recommend you get a film holder which can hold your negatives 100% as close to the glass as possible.

I'll show you the details of my tests below, but the conclusion is that the Reflecta is not bad: it delivers a sharps 3600dpi scan with effective ICE and reasonable auto-everything. The workflow is quite fast, since I just make 1 preview scan to position the cropping selection and then I make subsequent full resolution scans of all negatives using the button on the scanner. Switching to the next negative is just moving the negative holder to the next notch and pushing the button. Postprocessing is faster on the Reflecta than on the Epson, since I trust the auto settings more so I can automate postprocessing in Photoshop with an action (appy curve, boost saturation, unsharp mask, switch to 8 bit mode, save as jpeg). For the Epson I need to tweak the prescan settings for each individual negative. But the Epson has the advantage of being able to hold and scan all images on 2 negative strips in one go.

Below are the results of my tests. First we'll look at the features of the Reflecta Proscan 7200, then we'll compare the scanner to the Epson V500. All images are full resolution 3600dpi scans.

1. ICE test

The first image clearly shows how effective ICE is for dust removal. The second images shows how ICE helps to remove scratches. The Proscan has a very good ICE implementation.

2. GEM

This function is supposedly a software algorithm (whereas ICE is a hardware feature) to reduce film grain. The first image shows how it is effective in the highlights and midtones. The second image shows how GEM is introducing noise in the shadows instead of smoothing out grain. I guess this is because the Proscan has a limited dynamic range and loses quite a lot of shadow detail. (in both examples the top image is without GEM and the lower is has GEM enabled).

3. ROC

The ROC feature is supposed to restore colors on old negatives. I used it on a brand new negative and found it boosts saturation and contrast by eliminating more of the negative base color Whilst some people claim it causes image artifacts, I have not noticed any negative effects of using the ROC feature. (bottom image has ROC applied)

Comparison between the Epson V500 and the Reflecta Proscan 7200

1. Color

These are the color negative we'll use to compare both scanners. They are shot on Ektar 100 and should be sharp and low grain. (all other images below are full resolution 3600dpi crops)

The image below shows a lot more color noise in the shadows for the Proscan (under the breast). We also notice in the background that the Proscan captures more green in the shadows. This means the color balance of both images doesn't match, so the color noise can simply be a consequence of this.

In the image above you also see that the Epson applies more sharpening than the Proscan. In the image below I compensated the Proscan image using an unsharp mask filter. We now see similar sharpness, but of course even more pronounced noise.

The image below shows another crop where it is clear that the Proscan manages to get more shadow detail out of the negative then the Epson. (look at the top lef) This is somewhat contradictory to the more limited dynamic range (in the Epson software I have a lot of room to the left and right of the histogram whilst configuring the scanner settings, in the Proscan there is clipping on both sides even when the histogram sliders are at their maximum).

Let's switch to the other testimage to check for sharpness. In the small branches, you notice sharpness is just a tiny bit better on the Proscan than the Epson. But the difference is marginal, contrary to what my expected/hoped.

The color balance on both scans of this image is much more similar than in our previous testimage, so let's compare noise again in the picture below. We see the Epson now has more midrange noise (look in the blue band). The shadows in the black are clipped so there is no visible noise in both images. Looking at the green leaves on the left we notice again that more dark greens are detected by the Proscan than by the Epson, leading to the perception of more color noise.

2. Black & White

The black and white image we used for our comparison is shot on Tmax 100 and should be sharp and feature low grain. This is the image we used (all other images are full res crops at 3600dpi).

When we compare sharpness, the Proscan is again slightly sharper than the Epson. But not by much.

When we look at the shadow detail, this time the Epson wins from the Proscan. (look at the tree on the left) I guess this proves that the overall Dmax of the Epson is better than the Proscan, but the proscan has better separation in the low density areas of the negative (=shadow areas of the positive image), leading to more color in the shadows.

CONCLUSION

I regret buying this scanner. Not because it is bad, but because it is only marginally better than my Epson v500, which was only about €50 on eBay. The Proscan is slightly sharper, has better autoscan settings, has slightly better ICE and more color detail in the shadows. I was expecting a bigger difference, based on the reviews on http://www.filmscanner.info (I also bought the scanner on their http://www.scandig.com/site for a good price and hasslefree service). Maybe I want to try a Plustek OpticFilm 8200i. According to filmscanner.info this scanner has about the same measured resolution as the Proscan, but maybe the extra pixels will show a higher perceived resolution. You can't express everything in numbers.

An experiment with sprocket hole photography

2013-04-09T07:56:00.001-07:00

I decided to experiment with sprocket hole photography with my Rolleicord on a cheap Fuji c200 film. Pictures are scanned emulsion side down on the glass with an Epson v500.

No special 35mm adapter was used: I just inserted the film at the bottom and guided it through the pickup spool at the top. The film cassette was held more or less in place by two blocks of foam I cut for the purpose. I used the camera on the red setting to NOT prevent double exposures, since you can't use the exposure counter anymore. When the advance mechanism engaged whilst I was forwarding the film, I just put the lens cap on, pressed the shutter at 1/500 and continued winding until I was at the next frame.

How did I determine how much to wind on? I measured the diameter of the empty pickup spool and I estimated the diameter of the full pickup spool. By using some mathematics, I figured out I should start with 2.3 turns for the first picture, gradually lowering to 0.75 turns for the last picture. This turned out just fine, since I had enough space between all of my negatives. I once made the mistake of advancing the film twice, so I lost a picture there. Otherwise I would have had 20 negatives on the roll out of a theoretical maximum of 24. So my formula worked out fine.

I noticed the Epson v500 is much sharper when the negative is flat on the glass. If you scan with the emulsion side down you will also avoid problems with Newton rings. I didn't put a glass on top of the negative but just held down the edges of the negative with 2 small booklets (Berlitx pocket travel guides).

Next time, I will try scanning my normal negatives this way. Maybe I'll have to make a special purpose negative holder to hold the negatives to the scanner glass.

Regarding the sprocket hole photography: I noticed something had scratched the negatives. The Rolleicord never did this before with 120 film, so it will probably be a grain of salt in the film cassette.

Some pics are fun, others are useless (I was often wrong about the vertical versus horizontal framing...). It was nice to do this for a change, but I prefer regular photography.

More pictures (and 15Mpx full resolution downloads) in the Flickr album:
http://www.flickr.com/photos/kodel/sets/72157633200794551/

Mirror slap: highly overrated?

2013-01-12T04:10:00.001-08:00

Mirror slap is the effect in DSLR camera's where the image loses sharpness due to camera movement induced by the miror flipping up.
I tried to do some tests on the Nikon EM, which is supposed to have really bad mirror slap. I mounted the camera on a tripod, attached a 300mm telelens and took 3 consecutive pictures: one without mirror lockup, one with mirror lockup (accomplished on the Nikon EM by using the self-timer) and one with a modified mirror bumper.

Let me explain you the modified mirror bumper. When the mirror moves out of the way for the picture to be taken, it needs to stop at some point. The transition from the mirror being in movement to the miror being halted when it is folded against the groundglass is only possible when it decelerates. The deceleration creates a force (momentum) which will be counteracted by an acceleration of the camera in the opposite direction. This is what is inducing camera shake.
To minimise camera shake, we need to minimise the deceleration force on the mirror. To this end, a foam bumper is mounted at the top of the groundglass. The idea is that it will bring the mirror progressively to a halt. My modifified mirror bumper is a mirror bumper of double tickness, but modified in the form of a sawtooth. This causes the deceleration to start earlier with the resistance of the foam progressively increasing. The end result should be (theoretically) that the acceleration force the camera will be subjected to will be a lot less (about half).

The theory is nice, but the test results are not convincing: the 3 images (left: normal mirror slap, center: no mirror slap due to mirror lockup, right: reduced mirror slap with modified mirror bumper) show no difference whatsoever (check for yourself: the image inserted in this post is a 4800dpi enlargement of the center of the 3 negatives). Either this means my test was not valid and I need a sharper film, lens or scanner. Or it could be that mirror slap is highly overrated. Note that I tested for the worst possible conditions: light camera, heavily bouncing mirror, long telelens and a shutter speed of 1/15s to maximise mirror slap.
I think many people overrate mirror slap because of the camera shake they feel in their hands. Most of this camera shake is not having an effect on the image, since it occurs when the mirror flips back into it's resting position (where there is no foam to dampen the bouncing). Since the shutter curtain is already closed at this point, it is not influencing the picture.

A second, less scientific test I did was comparing my Nikon EM (light camera, bouncy mirror) with my Nikon F-301 (heavy camera, smooth mirror) for handheld available light shooting. In this case, I definitely had more sharp shots on 1/15th and even sometimes 1/8th when bracing myself with the F-301 than with the EM. But this can also be because the size and weight of the larger camera make it easier to hold it still, without miror slap having anything to do with it.

Pushing Portra 400 to 1600 iso

2012-12-28T04:31:00.001-08:00

Available light photography has become the domain of DSLR's with their high iso ratings. Yet, I wanted to try what film can do and I pushed a 35mm roll of Porta 400 with 2 stops to 1600 iso. I compensated in development by extending the normal 3min15 of the c41 process with 1 extra minute. The images are usable with a typical available light atmosphere. I got major color shifts, not only due to the pushing but also due to various indoor light sources.

high res pics:
http://www.flickr.com/photos/kodel/sets/72157632365305794
https://plus.google.com/photos/112874694436727534142/albums/5826962777139070129

Further underexposure does not yield good results. The shot below is exposed at 6400, so 2 stops under.

Only 1 stop underexposure is still usable, but quality suffers.

The biggest problem is a complete lack of blue. If you try to bump up the blue channel to get more or less normal colours lots of blue grain pops up. The blue grain also could be scanner noise. I would have to make a print in a darkroom to find out.

The shot below is one of the only shots that is lit with a bit of daylight through the window. The color shifts are much less severe and it looks more like a normal Portra shot.

All photos are taken at around maximum aperture of my Nikon 50mm 1.8 EM lens. This resulted in shutter speeds of around 1/30 second. This is where the challenge for available light photography lies: shooting moving subjects with critical focussing and low handheld speeds. Below is a shot where I took the time to focus correctly and could brace myself to hold the camera more steady. As a result, sharpness is much better than in the shot above.

It's amazing that a color film can be pushed to 1600 iso without too much problems. I wonder if we could even go further to 3200 or 6400 iso. I still prefer a good black and white shot on TMAX 3200, but since I have only 1 roll left in my fridge and Kodak discontinued production I'll have to decide between pushing Portra or using my DSLR at 6400 iso. I'm afraid that for most situations, the DSLR will win. The only area where the pushed Portra is better is in controlling the highlights. You can even retain detail in the flame of a candle. But the DSLR wins from a practical point of view: fast autofocus, higher shutter speeds, less noise and more detail.

What are your thoughts?